Commit Graph

7 Commits

Author SHA1 Message Date
Ward Fisher
771b959cad
Merge branch 'main' into jsonconvention.dmh 2022-05-17 13:24:53 -06:00
Ward Fisher
375e5adfe4
Merge branch 'main' into alwaysxarray.dmh 2022-05-17 13:23:19 -06:00
Dennis Heimbigner
444024a7be Merge branch 'master' into jsonconvention.dmh 2022-05-01 13:16:58 -06:00
Dennis Heimbigner
9f78be8bb8 Allow the read/write of JSON-valued Zarr attributes.
A number of other packages that read/write Zarr insert
attributes whose value is a dictionary containing specialized
information.  An example is the GDAL Driver convention (see
https://gdal.org/drivers/raster/zarr.html).

In order to handle such attributes, this PR enforces a special
convention. It applies to both pure Zarr an NCZarr format as
written by the netdf-c library.

The convention is as follows:

## Reading
Suppose an attribute is read from *.zattrs* and it has a JSON
value that is a a dictionary.  In this case, the JSON dictionary
is converted to a string value.  It then appears in the netcdf-c
API as if it is a character valued attribute of the same name,
and whose value is the "stringified" dictionary.

# Writing
Suppose an attribute is of type character and its *value* *looks like*
a JSON dictionary. In this case, it is parsed to JSON
and written as the value of the attribute in the NCZarr file.
Here the *value* is the concatenation of all the characters
in the attributes netcdf-c value.
The term "looks like" means that the *value*'s first character is
"{", its last value is "}", and it can be successfully parsed
by a JSON parser.

A test case, *nczarr_test/run_jsonconventions.sh* was also added.

## Misc. Unrelated Changes

1. Fix an error in nc_test4/tst_broken_files.c
2. Modify the internal JSON parser API.
3. Modify the nczarr_test/zisjson program is modified to support
   this convention.
2022-04-06 18:22:59 -06:00
Dennis Heimbigner
a5cae51efc Turn on the xarray convention for NCZarr format
re: https://github.com/pydata/xarray/issues/6374

As a result of a discussion about Xarray (see above issue),
I decided to turn on the xarray convention for NCZarr datasets
where possible so that xarray can read a larger set of nczarr
generated datasets.

This causes the following changes:
* If the user wants to generate a pure zarr file, then the mode "zarr" must be explicitly used; it is no longer the case that "mode=xarray" or mode="noxarray"
implies "mode=zarr".
* It is still the case that "mode=noxarray" will turn off the XArray convention.

The following conditions will cause ''_ARRAY_DIMENSIONS'' to not be written.
* The variable is not in the root group,
* Any dimension referenced by the variable is not in the root group.
2022-03-23 19:28:06 -06:00
Dennis Heimbigner
a83b7b9fe7 Clean up some ncdap tests.
This is in response to issue https://github.com/Unidata/netcdf-c/issues/2188#issuecomment-1072762308 although it does not fix that problem.
2022-03-18 21:01:40 -06:00
Tobias Kölling
b47f61bc25 added test for opening a trunceded file using in-memory open
This test currently fails.
When trying to open an in-memory file which is only available partially
(i.e. by accidental truncation), then the library should fail with an
error in stead of an assertation, such that user code can react on this.
2021-08-24 17:04:40 +02:00